TIME
Rightscom, EPICentre, and Book Industry Communication developed a testbed ‘engine’ to transform metadata from one format to another, via a generic format, to facilitate the cataloguing of e-books.
Full Project Title
Testbed for Interoperability of Metadata for E-Books
Summary
Recently there has been a rapid growth in e-books. The JISC e-Books working group has commissioned a series of studies examining the potential of e-books in higher/further education and the related barriers to take-up. One of the barriers identified was the lack of standardised e-book catalogue records and the lack of interoperability between different e-book metadata records. In an ideal world it should be easy to catalogue e-books and then reuse the metadata in other user environments, e.g. in a virtual learning environment (VLE). This is difficult if publishers and libraries use different metadata standards and the resulting records are not interoperable.
The TIME project set out to develop a solution to this interoperability problem. Rightscom, EPICentre, and Book Industry Communication (BIC) collaborated to develop a testbed ‘transformation engine’. The engine transforms e-book metadata from a variety of formats into a highly granular and meaningful generic format. Once in the generic format, it can then be output into the same range of formats. Input/output formats for the project included ONIX for Books, Dublin Core, MARC, and IEEE LOM. So if the user has e-book metadata in ONIX format, these records could be transformed into MARC for a library catalogue or IEEE LOM for a VLE.
The testbed is based on Rightscom’s Core Ontologyx Architecture, an ontology based methodology to support syntactic and semantic mapping between different metadata schemas. It is designed so it can be expanded to include any number of input and output formats. During the project 1,900 input records from four publishers were transformed and the resulting output records evaluated. An OAI-PMH interface was also developed to allow harvesting from the testbed. At this point, it’s only a testbed, and more work would be needed to develop it into a fully functional system.
The project raised a number of interesting issues about metadata transformation for e-books. It demonstrated that the ‘hub’ of a transformation engine needs to be as rich as all of the ‘spokes’ put together, and the mappings need to preserve all their semantics in the ‘hub’. It also demonstrated that some schemas work out better than others. ONIX and MARC worked out well as input formats where unqualified Dublin Core did not (no controlled values). However, it is possible to generate Dublin Core from ONIX and MARC. Numerous issues related to data quality were also identified.
Outcomes
The project will explore with JISC the potential of the transformation engine for e-book cataloguing. There may be potential for union catalogues to use the engine and then make the records available to member libraries.
Outputs
Demonstrator
The core project deliverable is the testbed transformation engine and supporting installation and user documentation. Contact Balviar Notay at JISC for further information.
Reports
The project’s Final Report (PDF) describes how the requirements were specified and validated and the testbed was developed. It fully documents the architecture, transformation methodology, and technical conclusions.
Presentations
Further information